[nSLUG] Ubuntu crash with scp.

Rich budman85 at eastlink.ca
Thu Nov 8 00:09:09 AST 2007

David Payne wrote:
> On Wed, Nov 07, 2007 at 08:22:31PM -0400, George N. White III wrote:
>> On 11/7/07, David Payne <david at payneful.ca> wrote:
>> This week, my first suspicion would be power fluctuations.  Are both
>> systems and the network all on UPS's?
>> In the past I have similar issues with unix.  Tracking down the
>> problem involved way too much work: finding a way to reproduce the
>> problem and then finding space (this was when 9G was the largest SCSI
>> disk you could get) for huge process logs and network packets dumps
>> and waiding thru them to find the smoking gun.
> Hi,
> I'm afraid it's not power fluctuations.  I am on an 
> internship in Markham, Ontario.  I'm on this list because 
> I'm from Nova Scotia and I'm going to come back after the 
> internship is over.  They didn't have the same power 
> problems up here.  The systems are not on a UPS so that 
> still could be the problem.

Have you tried a strace to see if its something local?
       strace -f "your scp cmd" -o logfile
       if its stuck, either as the user or root, run
          strace -p pid#   -o logfile

       to see if its hung or waiting 
Does it occur with all hosts or only one host?
Try other hosts to see if its your problem.

Are you running a trusted ssh connection or do you update the keys in .ssh ?
Maybe the known-hosts  has an issue.

Network connection - anything new?  Can ssh or ftp to the machine your 
sending to?
What type of controller do you have?
Maybe the card it auto negotiating under load conditions.
Do you have a router - any port blocking or timed constraints? (after 
such and such an hour ports are blocked)

How many drives do you have?
Any interrupts sharing with the network card - check /proc/interrupts
Check the ioports as well.

Run ifconfig on your nic to see what interrupt you are using.
Maybe there is a memory conflict -  base address or buffer area is 
Do you have anything else that is running using the ide (or is it 
another type) where the file is you are sending?
How large is the file you are sending - maybe you are maxing buffers.

Can you give us a brief overview of what's active on the machine when 
you run into the lockup?


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://nslug.ns.ca/mailman/private/nslug/attachments/20071108/3ce93644/attachment.html>

More information about the nSLUG mailing list